JCK : JCK6.0 b19
J2SE : FAIL - mustang b76
Platform[s] : FAIL - Windows XP Pro, Windows XP Home
switch/Mode : FAIL - -Xconcgc | -Xincgc | -Xnoincgc
Reproducable : *NOT ALWAYS*


test api/javax_sound/sampled/SourceDataLine/index.html#SourceDataLine hangs on Windows platform if any GC option is specified. *It seems to be some kind of race condition*.

This test is reproducable on some of our PCs. On the workstation (Windows XP Home) there this bug was originally found attached test sample doesn't hang, but JCK test still hangs. On some of our workstations (Windows XP Pro) test hangs also on b75, on some - not.

Steps to reproduce:

1. get attached class Sound.java
2. compile it using JDK b76
3. Comment line 61 (it contains *System.out.println("Hello, World!!!");*)
4. run it on b75. On b75 it returns (for my PC):


C:\tests\sound>Z:/Links/stt/jdk_promotions/JDK6.0/b75/binaries/windows-i586/jdk1.6.0/bin/java -cp . Sound
testing a SourceDataLine with format: PCM_SIGNED 44100.0 Hz, 16 bit, stereo, 4 bytes/frame, little-endian
testing a SourceDataLine with format: PCM_SIGNED 44100.0 Hz, 16 bit, stereo, 4 bytes/frame, little-endian
testing a SourceDataLine with format: PCM_SIGNED 44100.0 Hz, 16 bit, stereo, 4 bytes/frame, little-endian
OKAY

5. Run it on b76. It returns first two rows then hangs.

6. Uncomment line 61 then compile and run on b76 again. Test passes.